Can you elaborate on how you looked at the raw data? Did you enable features that show encoded characters (e.g., a line end would show up as \r\n for windows text files.)

That appears to be utf-8 encoding (perhaps of an em-dash or something else which 'looks' normal). Is it in the expected encoding format?

One way to debug this would be to split the file-- remove the first five lines, and see if it reads in. If so, the issue is being caused by something in the first five lines (etc.) With csvs, that can be the result of richer text in the header rows. It's not elegant, but it might help you learn more about a problem like this.